includes/job/JobQueueGroup.php

   1 <?php
   2 /**
   3  * Job queue base code.
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  * @author Aaron Schulz
  22  */
  23
  24 /**
  25  * Class to handle enqueueing of background jobs
  26  *
  27  * @ingroup JobQueue
  28  * @since 1.21
  29  */
  30 class JobQueueGroup {
  31         /** @var array */
  32         protected static $instances = array();
  33
  34         /** @var ProcessCacheLRU */
  35         protected $cache;
  36
  37         /** @var string Wiki ID */
  38         protected $wiki;
  39
  40         /** @var array Map of (bucket => (queue => JobQueue, types => list of types) */
  41         protected $coalescedQueues;
  42
  43         const TYPE_DEFAULT = 1; // integer; jobs popped by default
  44         const TYPE_ANY = 2; // integer; any job
  45
  46         const USE_CACHE = 1; // integer; use process or persistent cache
  47         const USE_PRIORITY = 2; // integer; respect deprioritization
  48
  49         const PROC_CACHE_TTL = 15; // integer; seconds
  50
  51         const CACHE_VERSION = 1; // integer; cache version
  52
  53         /**
  54          * @param string $wiki Wiki ID
  55          */
  56         protected function __construct( $wiki ) {
  57                 $this->wiki = $wiki;
  58                 $this->cache = new ProcessCacheLRU( 10 );
  59         }
  60
  61         /**
  62          * @param bool|string $wiki Wiki ID
  63          * @return JobQueueGroup
  64          */
  65         public static function singleton( $wiki = false ) {
  66                 $wiki = ( $wiki === false ) ? wfWikiID() : $wiki;
  67                 if ( !isset( self::$instances[$wiki] ) ) {
  68                         self::$instances[$wiki] = new self( $wiki );
  69                 }
  70
  71                 return self::$instances[$wiki];
  72         }
  73
  74         /**
  75          * Destroy the singleton instances
  76          *
  77          * @return void
  78          */
  79         public static function destroySingletons() {
  80                 self::$instances = array();
  81         }
  82
  83         /**
  84          * Get the job queue object for a given queue type
  85          *
  86          * @param string $type
  87          * @return JobQueue
  88          */
  89         public function get( $type ) {
  90                 global $wgJobTypeConf;
  91
  92                 $conf = array( 'wiki' => $this->wiki, 'type' => $type );
  93                 if ( isset( $wgJobTypeConf[$type] ) ) {
  94                         $conf = $conf + $wgJobTypeConf[$type];
  95                 } else {
  96                         $conf = $conf + $wgJobTypeConf['default'];
  97                 }
  98
  99                 return JobQueue::factory( $conf );
 100         }
 101
 102         /**
 103          * Insert jobs into the respective queues of with the belong.
 104          *
 105          * This inserts the jobs into the queue specified by $wgJobTypeConf
 106          * and updates the aggregate job queue information cache as needed.
 107          *
 108          * @param Job|array $jobs A single Job or a list of Jobs
 109          * @throws MWException
 110          * @return bool
 111          */
 112         public function push( $jobs ) {
 113                 $jobs = is_array( $jobs ) ? $jobs : array( $jobs );
 114
 115                 $jobsByType = array(); // (job type => list of jobs)
 116                 foreach ( $jobs as $job ) {
 117                         if ( $job instanceof Job ) {
 118                                 $jobsByType[$job->getType()][] = $job;
 119                         } else {
 120                                 throw new MWException( "Attempted to push a non-Job object into a queue." );
 121                         }
 122                 }
 123
 124                 $ok = true;
 125                 foreach ( $jobsByType as $type => $jobs ) {
 126                         if ( $this->get( $type )->push( $jobs ) ) {
 127                                 JobQueueAggregator::singleton()->notifyQueueNonEmpty( $this->wiki, $type );
 128                         } else {
 129                                 $ok = false;
 130                         }
 131                 }
 132
 133                 if ( $this->cache->has( 'queues-ready', 'list' ) ) {
 134                         $list = $this->cache->get( 'queues-ready', 'list' );
 135                         if ( count( array_diff( array_keys( $jobsByType ), $list ) ) ) {
 136                                 $this->cache->clear( 'queues-ready' );
 137                         }
 138                 }
 139
 140                 return $ok;
 141         }
 142
 143         /**
 144          * Pop a job off one of the job queues
 145          *
 146          * This pops a job off a queue as specified by $wgJobTypeConf and
 147          * updates the aggregate job queue information cache as needed.
 148          *
 149          * @param int|string $qtype JobQueueGroup::TYPE_DEFAULT or type string
 150          * @param int $flags Bitfield of JobQueueGroup::USE_* constants
 151          * @return Job|bool Returns false on failure
 152          */
 153         public function pop( $qtype = self::TYPE_DEFAULT, $flags = 0 ) {
 154                 if ( is_string( $qtype ) ) { // specific job type
 155                         if ( ( $flags & self::USE_PRIORITY ) && $this->isQueueDeprioritized( $qtype ) ) {
 156                                 return false; // back off
 157                         }
 158                         $job = $this->get( $qtype )->pop();
 159                         if ( !$job ) {
 160                                 JobQueueAggregator::singleton()->notifyQueueEmpty( $this->wiki, $qtype );
 161                         }
 162
 163                         return $job;
 164                 } else { // any job in the "default" jobs types
 165                         if ( $flags & self::USE_CACHE ) {
 166                                 if ( !$this->cache->has( 'queues-ready', 'list', self::PROC_CACHE_TTL ) ) {
 167                                         $this->cache->set( 'queues-ready', 'list', $this->getQueuesWithJobs() );
 168                                 }
 169                                 $types = $this->cache->get( 'queues-ready', 'list' );
 170                         } else {
 171                                 $types = $this->getQueuesWithJobs();
 172                         }
 173
 174                         if ( $qtype == self::TYPE_DEFAULT ) {
 175                                 $types = array_intersect( $types, $this->getDefaultQueueTypes() );
 176                         }
 177                         shuffle( $types ); // avoid starvation
 178
 179                         foreach ( $types as $type ) { // for each queue...
 180                                 if ( ( $flags & self::USE_PRIORITY ) && $this->isQueueDeprioritized( $type ) ) {
 181                                         continue; // back off
 182                                 }
 183                                 $job = $this->get( $type )->pop();
 184                                 if ( $job ) { // found
 185                                         return $job;
 186                                 } else { // not found
 187                                         JobQueueAggregator::singleton()->notifyQueueEmpty( $this->wiki, $type );
 188                                         $this->cache->clear( 'queues-ready' );
 189                                 }
 190                         }
 191
 192                         return false; // no jobs found
 193                 }
 194         }
 195
 196         /**
 197          * Acknowledge that a job was completed
 198          *
 199          * @param Job $job
 200          * @return bool
 201          */
 202         public function ack( Job $job ) {
 203                 return $this->get( $job->getType() )->ack( $job );
 204         }
 205
 206         /**
 207          * Register the "root job" of a given job into the queue for de-duplication.
 208          * This should only be called right *after* all the new jobs have been inserted.
 209          *
 210          * @param Job $job
 211          * @return bool
 212          */
 213         public function deduplicateRootJob( Job $job ) {
 214                 return $this->get( $job->getType() )->deduplicateRootJob( $job );
 215         }
 216
 217         /**
 218          * Wait for any slaves or backup queue servers to catch up.
 219          *
 220          * This does nothing for certain queue classes.
 221          *
 222          * @return void
 223          * @throws MWException
 224          */
 225         public function waitForBackups() {
 226                 global $wgJobTypeConf;
 227
 228                 wfProfileIn( __METHOD__ );
 229                 // Try to avoid doing this more than once per queue storage medium
 230                 foreach ( $wgJobTypeConf as $type => $conf ) {
 231                         $this->get( $type )->waitForBackups();
 232                 }
 233                 wfProfileOut( __METHOD__ );
 234         }
 235
 236         /**
 237          * Get the list of queue types
 238          *
 239          * @return array List of strings
 240          */
 241         public function getQueueTypes() {
 242                 return array_keys( $this->getCachedConfigVar( 'wgJobClasses' ) );
 243         }
 244
 245         /**
 246          * Get the list of default queue types
 247          *
 248          * @return array List of strings
 249          */
 250         public function getDefaultQueueTypes() {
 251                 global $wgJobTypesExcludedFromDefaultQueue;
 252
 253                 return array_diff( $this->getQueueTypes(), $wgJobTypesExcludedFromDefaultQueue );
 254         }
 255
 256         /**
 257          * Get the list of job types that have non-empty queues
 258          *
 259          * @return array List of job types that have non-empty queues
 260          */
 261         public function getQueuesWithJobs() {
 262                 $types = array();
 263                 foreach ( $this->getCoalescedQueues() as $info ) {
 264                         $nonEmpty = $info['queue']->getSiblingQueuesWithJobs( $this->getQueueTypes() );
 265                         if ( is_array( $nonEmpty ) ) { // batching features supported
 266                                 $types = array_merge( $types, $nonEmpty );
 267                         } else { // we have to go through the queues in the bucket one-by-one
 268                                 foreach ( $info['types'] as $type ) {
 269                                         if ( !$this->get( $type )->isEmpty() ) {
 270                                                 $types[] = $type;
 271                                         }
 272                                 }
 273                         }
 274                 }
 275
 276                 return $types;
 277         }
 278
 279         /**
 280          * Get the size of the queus for a list of job types
 281          *
 282          * @return array Map of (job type => size)
 283          */
 284         public function getQueueSizes() {
 285                 $sizeMap = array();
 286                 foreach ( $this->getCoalescedQueues() as $info ) {
 287                         $sizes = $info['queue']->getSiblingQueueSizes( $this->getQueueTypes() );
 288                         if ( is_array( $sizes ) ) { // batching features supported
 289                                 $sizeMap = $sizeMap + $sizes;
 290                         } else { // we have to go through the queues in the bucket one-by-one
 291                                 foreach ( $info['types'] as $type ) {
 292                                         $sizeMap[$type] = $this->get( $type )->getSize();
 293                                 }
 294                         }
 295                 }
 296
 297                 return $sizeMap;
 298         }
 299
 300         /**
 301          * @return array
 302          */
 303         protected function getCoalescedQueues() {
 304                 global $wgJobTypeConf;
 305
 306                 if ( $this->coalescedQueues === null ) {
 307                         $this->coalescedQueues = array();
 308                         foreach ( $wgJobTypeConf as $type => $conf ) {
 309                                 $queue = JobQueue::factory(
 310                                         array( 'wiki' => $this->wiki, 'type' => 'null' ) + $conf );
 311                                 $loc = $queue->getCoalesceLocationInternal();
 312                                 if ( !isset( $this->coalescedQueues[$loc] ) ) {
 313                                         $this->coalescedQueues[$loc]['queue'] = $queue;
 314                                         $this->coalescedQueues[$loc]['types'] = array();
 315                                 }
 316                                 if ( $type === 'default' ) {
 317                                         $this->coalescedQueues[$loc]['types'] = array_merge(
 318                                                 $this->coalescedQueues[$loc]['types'],
 319                                                 array_diff( $this->getQueueTypes(), array_keys( $wgJobTypeConf ) )
 320                                         );
 321                                 } else {
 322                                         $this->coalescedQueues[$loc]['types'][] = $type;
 323                                 }
 324                         }
 325                 }
 326
 327                 return $this->coalescedQueues;
 328         }
 329
 330         /**
 331          * Check if jobs should not be popped of a queue right now.
 332          * This is only used for performance, such as to avoid spamming
 333          * the queue with many sub-jobs before they actually get run.
 334          *
 335          * @param string $type
 336          * @return bool
 337          */
 338         public function isQueueDeprioritized( $type ) {
 339                 if ( $this->cache->has( 'isDeprioritized', $type, 5 ) ) {
 340                         return $this->cache->get( 'isDeprioritized', $type );
 341                 }
 342                 if ( $type === 'refreshLinks2' ) {
 343                         // Don't keep converting refreshLinks2 => refreshLinks jobs if the
 344                         // later jobs have not been done yet. This helps throttle queue spam.
 345                         $deprioritized = !$this->get( 'refreshLinks' )->isEmpty();
 346                         $this->cache->set( 'isDeprioritized', $type, $deprioritized );
 347
 348                         return $deprioritized;
 349                 }
 350
 351                 return false;
 352         }
 353
 354         /**
 355          * Execute any due periodic queue maintenance tasks for all queues.
 356          *
 357          * A task is "due" if the time ellapsed since the last run is greater than
 358          * the defined run period. Concurrent calls to this function will cause tasks
 359          * to be attempted twice, so they may need their own methods of mutual exclusion.
 360          *
 361          * @return int Number of tasks run
 362          */
 363         public function executeReadyPeriodicTasks() {
 364                 global $wgMemc;
 365
 366                 list( $db, $prefix ) = wfSplitWikiID( $this->wiki );
 367                 $key = wfForeignMemcKey( $db, $prefix, 'jobqueuegroup', 'taskruns', 'v1' );
 368                 $lastRuns = $wgMemc->get( $key ); // (queue => task => UNIX timestamp)
 369
 370                 $count = 0;
 371                 $tasksRun = array(); // (queue => task => UNIX timestamp)
 372                 foreach ( $this->getQueueTypes() as $type ) {
 373                         $queue = $this->get( $type );
 374                         foreach ( $queue->getPeriodicTasks() as $task => $definition ) {
 375                                 if ( $definition['period'] <= 0 ) {
 376                                         continue; // disabled
 377                                 } elseif ( !isset( $lastRuns[$type][$task] )
 378                                         || $lastRuns[$type][$task] < ( time() - $definition['period'] )
 379                                 ) {
 380                                         try {
 381                                                 if ( call_user_func( $definition['callback'] ) !== null ) {
 382                                                         $tasksRun[$type][$task] = time();
 383                                                         ++$count;
 384                                                 }
 385                                         } catch ( JobQueueError $e ) {
 386                                                 MWExceptionHandler::logException( $e );
 387                                         }
 388                                 }
 389                         }
 390                 }
 391
 392                 $wgMemc->merge( $key, function ( $cache, $key, $lastRuns ) use ( $tasksRun ) {
 393                         if ( is_array( $lastRuns ) ) {
 394                                 foreach ( $tasksRun as $type => $tasks ) {
 395                                         foreach ( $tasks as $task => $timestamp ) {
 396                                                 if ( !isset( $lastRuns[$type][$task] )
 397                                                         || $timestamp > $lastRuns[$type][$task]
 398                                                 ) {
 399                                                         $lastRuns[$type][$task] = $timestamp;
 400                                                 }
 401                                         }
 402                                 }
 403                         } else {
 404                                 $lastRuns = $tasksRun;
 405                         }
 406
 407                         return $lastRuns;
 408                 } );
 409
 410                 return $count;
 411         }
 412
 413         /**
 414          * @param $name string
 415          * @return mixed
 416          */
 417         private function getCachedConfigVar( $name ) {
 418                 global $wgConf, $wgMemc;
 419
 420                 if ( $this->wiki === wfWikiID() ) {
 421                         return $GLOBALS[$name]; // common case
 422                 } else {
 423                         list( $db, $prefix ) = wfSplitWikiID( $this->wiki );
 424                         $key = wfForeignMemcKey( $db, $prefix, 'configvalue', $name );
 425                         $value = $wgMemc->get( $key ); // ('v' => ...) or false
 426                         if ( is_array( $value ) ) {
 427                                 return $value['v'];
 428                         } else {
 429                                 $value = $wgConf->getConfig( $this->wiki, $name );
 430                                 $wgMemc->set( $key, array( 'v' => $value ), 86400 + mt_rand( 0, 86400 ) );
 431
 432                                 return $value;
 433                         }
 434                 }
 435         }
 436 }